An accurate IR land surface emissivity (LSE) model is essential for IR window channels simulation. In this work, we investigated the potential impacts of two IR land surface emissivity models on GFS forecast. The two emissivity models are recently connected with CRTM model. The first one is the NASA/LARC IR land emissivity model retrieved from IASI observation and the second one is UW-RTTOV IR emissivity module. The major difference and variance of the two emissivity models are over desert regions, where UW emissivity is smaller than IASI emissivity for about 0.01 – 0.02 in both summer and winter times for most spectral bands from 4 to 13 μm. The CRTM simulations at SEVIRI four surface sensitive channels are improved by using the two new emissivity models. There is little change for the simulation over water vapor and CO2 bands.

Three one-month GFS forecast experiments are performed by using the two new IR emissivity models in GSI data assimilation system, and compared with current CRTM look-up-table IR emissivity. Small positive impacts were found by using the two new IR land surface emissivity models. There is notable positive impact on GFS forecast from 4 to 7 days when using IASI emissivity model. The total assimilated observation numbers for IR sensors are increased, and there is very little change of the assimilation numbers for microwave sensors. In general for IR sensors, like IASI and AIRS, the observation simulation numbers show bigger increase over IR window bands, but smaller decrease over WV and CO2 bands. More detail diagnostic study of the experiment results will be given at the conference.